CSF cellranger QC

We are going to perform a quality control (QC) analysis of the mapping results obtained by running cellranger version 7.0.0.

We will pull together all the libraries from all the CSF subprojects.

## [1] "CSF_01"
## [1] "4608"
## [1] "4839"
## [1] "5700"
## [1] "5792"
## [1] "5929"
## [1] "CSF_02"
## [1] "7921"
## [1] "7974"

1 Gene Expression QC

We will start by showing the three most relevant metrics (number of reads, estimated number of recovered cells, fraction of reads in cells, mean reads per cell, fraction of reads mapped to exonic reads, and median genes per cell) obtained by cellranger for each of the working libraries. This information will give us an idea of the quality of the experiment as well as the sequencing and the mapping steps.

GEX QC metrics
cellranger v 7.0.0
Subproject	GemID	Cells	Median UMI counts per cell	Median genes per cell	Median reads per cell	Total genes detected	Number of reads
CSF_01	4608	3984	1623	858	19472	22942	266.97M
CSF_01	4839	3504	2964	1303	41738	22837	306.87M
CSF_01	5700	8989	2310	1049	16548	23935	298.90M
CSF_01	5792	7602	3659	1517	22742	25079	303.81M
CSF_01	5929	6962	3095	1319	16182	24309	261.68M
CSF_02	7921	4813	3998	1615	30334	23910	315.94M
CSF_02	7974	10196	3454	1565	12189	25758	231.41M

1.1 Mapping QC

Next, we will check the quality of the mapping step performed by cellranger 7.0.0 across libraries. To do so, we will compare the percentage of reads mapped to the genome, and within these mapped reads, the amount of reads mapped to intergenic regions, intronic and exonic regions. We aim to obtain libraries with a high percentage of confidently mapped reads, and specially a high percentage of exonic reads, which correspond with gene expression or RNAs. The reads mapping to intergenic regions suggest contamination of ambient DNA, whereas reads mapping to intronic regions may come from pre-mRNAs or mature spliced isoforms that retain certain introns.

## [1] "Confidently_mapped_to_genome"
## [1] "Confidently_mapped_to_intergenic_regions"
## [1] "Confidently_mapped_to_intronic_regions"
## [1] "Confidently_mapped_to_exonic_regions"
## [1] "Confidently_mapped_antisense"
## [1] "Confidently_mapped_to_transcriptome"

1.2 Sequencing saturation and depth

After assessing mapped reads, it is important to test which is the sequencing saturation and depth for each library. The sequencing saturation is dependent on the library complexity and sequencing depth. The library complexity is the total number of different transcripts present in the library and it varies between the cell types/tissues, whereas the sequencing depth is the number of paired reads per cell. For this reason, we will plot the number of detected genes as a function of depth (sequenced reads). As sequencing depth increases, more genes are detected, but this function reaches a plateau, whereby more sequenced reads does not result in more detected genes; therefore, at this point we assure we sequenced until saturation. More specifically, the sequencing saturation the fraction of confidently mapped, valid cell-barcode, valid UMI reads that had a non-unique (cell-barcode, UMI, gene).

Back to top

---

2 VDJ-T QC

We will start by showing the three most relevant metrics (number of reads, estimated number of recovered cells, fraction of reads in cells, mean reads per cell, fraction of reads mapped to any V(D)J gene, and cells with productive V-J Spanning Pair) obtained by cellranger for each of the working libraries. This information will give us an idea of the quality of the experiment as well as the sequencing and the mapping steps.

VDJ-T QC metrics
cellranger v 7.0.0
Subproject	GemID	Number of reads	Estimated number of cells	Fraction reads in cells	Mean reads per cell	Reads mapped to any V D J gene	Cells with productive V-J spanning pair
CSF_01	4608	13071011	815	34.80	0.02M	38.45	58.65
CSF_01	4839	15751069	1912	77.62	0.01M	72.19	77.62
CSF_01	5700	14127176	5719	62.66	0.00M	51.65	64.84
CSF_01	5792	18090339	2729	43.33	0.01M	31.17	69.11
CSF_01	5929	14573775	4297	52.51	0.00M	82.88	80.78

2.1 Mapping QC

Next, we will check the quality of the V(D)J mapping step performed by cellranger 7.0.0 across libraries. To do so, we will compare the percentage of reads mapped to any germline V(D)J gene segment, and within these mapped reads, the amount of reads mapped TRA and TRB germline gene segment.

2.2 V(D)J Expression

Here, we will assess the median number of UMIs assigned to a TRA/TRB contig per cell. Low values for any of the two parameters can indicate cells with extremely low TRA/TRB expression or poor cell quality, among others.

VDJ-T expression
cellranger v 7.0.0
GemID	Median_TRA_UMIs_per_Cell	Median_TRB_UMIs_per_Cell
4608	2	4
4839	3	7
5700	3	8
5792	3	7
5929	5	10

2.3 V(D)J Annotation

Now, we will check the V(D)J annotation for the studied samples. To better interpret the obtained results, we will consider the information given in the cellranger web summary file. We will assess the fraction of cell-associated barcodes (with at least…), that are the following ones:

• Cells With TRA/TRB Contig: one TRA/TRB contig annotated as a full or partial V(D)J gene.

• Cells With CDR3-annotated TRA/TRB Contig: one TRA/TRB contig where a CDR3 was detected.

• Cells With Productive TRA/TRB Contig: one contig that spans the 5’ end of the V region to the 3’ end of the J region for TRA/TRB, has a start codon in the expected part of the V sequence, has an in-frame CDR3, and has no stop codons in the aligned V-J region.

• Cells With Productive V-J Spanning Pair: one productive contig for each chain of the receptor pair. As well as the correspondent the number of cells with productive V-J Spanning Pair.

For all thre previous parameters, low values can indicate poor cell quality, low yield from the RT reaction, poor specificity of the V(D)J enrichment. Moreover, we will also check:

• Paired Clonotype Diversit Effective diversity of the paired clonotypes. It is computed as the Inverse Simpson Index of the clonotype frequencies. A value of 1 indicates a minimally diverse sample - only one distinct clonotype was detected, whereas a value equal to the estimated number of cells indicates a maximally diverse sample.

V(D)J annotation
cellranger v 7.0.0
GEM ID	Estimated Number of Recovered Cells	Productive V-J Spanning Pair		Paired Clonotype Diversity	Productive contig
		Fraction	Cells		TRA	TRB
4608	815	58.65	478	109.05	65.28	93.37
4839	1912	77.62	1484	457.31	80.65	96.97
5700	5719	64.84	3708	3103.42	68.86	95.98
5792	2729	69.11	1886	998.72	72.63	96.48
5929	4297	80.78	3471	648.62	83.17	97.60

Back to top

---

3 VDJ-B QC

We will start by showing the three most relevant metrics (number of reads, estimated number of recovered cells, fraction of reads in cells, mean reads per cell, fraction of reads mapped to any V(D)J gene, and cells with productive V-J Spanning Pair) obtained by cellranger for each of the working libraries. This information will give us an idea of the quality of the experiment as well as the sequencing and the mapping steps.

BCR-V(D)J QC metrics
cellranger v 7.0.0
GEM ID	Number of Reads	Estimated Number of Recovered Cells	Fraction of Reads in Cells	Mean Reads per Cell	Fraction of Reads Mapped to any VDJ gene	Cells With Productive V-J Spanning Pair
4839	13.55M	13	48.3%	1042457	47.6%	13
5929	13.87M	414	92.0%	33499	90.5%	387

3.1 Mapping QC

Next, we will check the quality of the V(D)J mapping step performed by cellranger 7.0.0 across libraries. To do so, we will compare the percentage of reads mapped to any germline V(D)J gene segment, and within these mapped reads, the amount of reads mapped IGH, IGK and IGL germline gene segment.

3.2 V(D)J Expression

Here, we will assess the median number of UMIs assigned to a IGH/IGK/IGL contig per cell. Low values for any of the three parameters can indicate cells with extremely low IGH/IGK/IGL expression or poor cell quality, among others.

V(D)J expression
cellranger v 7.0.0
GEM ID	Median IGH UMIs per Cell	Median IGK UMIs per Cell	Median IGL UMIs per Cell
4839	1995.0	7406	85
5929	2185.5	4343	5830

3.3 V(D)J Annotation

Now, we will check the V(D)J annotation for the studied samples. To better interpret the obtained results, we will consider the information given in the cellranger web summary file. We will assess the fraction of cell-associated barcodes (with at least…), that are the following ones:

• Cells With IGH/IGK/IGL Contig: one IGH/IGK/IGL contig annotated as a full or partial V(D)J gene.

• Cells With CDR3-annotated IGH/IGK/IGL Contig: one IGH/IGK/IGL contig where a CDR3 was detected.

• Cells With Productive IGH/IGK/IGL Contig: one contig that spans the 5’ end of the V region to the 3’ end of the J region for IGH/IGK/IGL, has a start codon in the expected part of the V sequence, has an in-frame CDR3, and has no stop codons in the aligned V-J region.

• Cells With Productive V-J Spanning Pair: one productive contig for each chain of the receptor pair. As well as the correspondent the number of cells with productive V-J Spanning Pair.

For all thre previous parameters, low values can indicate poor cell quality, low yield from the RT reaction, poor specificity of the V(D)J enrichment. Moreover, we will also check:

• Paired Clonotype Diversit Effective diversity of the paired clonotypes. It is computed as the Inverse Simpson Index of the clonotype frequencies. A value of 1 indicates a minimally diverse sample - only one distinct clonotype was detected, whereas a value equal to the estimated number of cells indicates a maximally diverse sample.

V(D)J annotation
cellranger v 7.0.0
Subproject	GEM ID	Productive contig			Productive V-J Spanning Pair			Cells_with_productive_V-J_spanning_pair	Estimated Number of Recovered Cells	Paired Clonotype Diversity	Reads_mapped_to_IGH	Reads_mapped_to_IGK	Reads_mapped_to_IGL
		IGH	IGK	IGL	Fraction IGK IGH Pair	Fraction IGL IGH Pair	Cells
CSF_01	4839	100.00	61.54	38.46	61.54	38.46	13	100.00	13	13.00	21.08	24.58	1.83
CSF_01	5929	93.48	67.39	33.09	63.04	30.92	387	93.48	414	50.95	33.41	38.64	18.40

4 Data overview

## [1] "Libraries metadata"

##    project subproject gem_id library_id library_name library_barcode    hashing
## 1      CSF     CSF_01   4608     276966     4608_GEX          AZ8142 not_hashed
## 2      CSF     CSF_01   4839     276967     4839_GEX          AZ8143 not_hashed
## 3      CSF     CSF_01   5929     276968     5929_GEX          AZ8144 not_hashed
## 4      CSF     CSF_01   5700     276969     5700_GEX          AZ8145 not_hashed
## 5      CSF     CSF_01   5792     276970     5792_GEX          AZ8146 not_hashed
## 6      CSF     CSF_01   4608     277405     4608_TCR          AZ8390 not_hashed
## 7      CSF     CSF_01   4839     277406     4839_TCR          AZ8391 not_hashed
## 8      CSF     CSF_01   5929     277407     5929_TCR          AZ8392 not_hashed
## 9      CSF     CSF_01   5700     277408     5700_TCR          AZ8393 not_hashed
## 10     CSF     CSF_01   5792     277409     5792_TCR          AZ8394 not_hashed
## 11     CSF     CSF_01   4839     277410     4839_BCR          AZ8396 not_hashed
## 12     CSF     CSF_01   5929     277411     5929_BCR          AZ8397 not_hashed
## 13     CSF     CSF_02   7921     277387     7921_GEX          AZ7864 not_hashed
## 14     CSF     CSF_02   7974     277388     7974_GEX          AZ7865 not_hashed
##     type donor_id  wet_lab
## 1   cDNA 4608_GEX 4608_GEX
## 2   cDNA 4839_GEX 4839_GEX
## 3   cDNA 5929_GEX 5929_GEX
## 4   cDNA 5700_GEX 5700_GEX
## 5   cDNA 5792_GEX 5792_GEX
## 6  VDJ-T 4608_TCR 4608_TCR
## 7  VDJ-T 4839_TCR 4839_TCR
## 8  VDJ-T 5929_TCR 5929_TCR
## 9  VDJ-T 5700_TCR 5700_TCR
## 10 VDJ-T 5792_TCR 5792_TCR
## 11 VDJ-B 4839_BCR 4839_BCR
## 12 VDJ-B 5929_BCR 5929_BCR
## 13  cDNA 7921_GEX 7921_GEX
## 14  cDNA 7974_GEX 7974_GEX

## [1] "GEX QC summary table"

## # A tibble: 7 x 26
##   Subproject GemID Cells Confi~1 Media~2 Media~3 Media~4 Total~5 Numbe~6 Numbe~7
##   <chr>      <chr> <dbl>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>
## 1 CSF_01     4608   3984    85.8    1623     858   19472   22942  2.67e8       0
## 2 CSF_01     4839   3504    93.3    2964    1303   41738   22837  3.07e8       0
## 3 CSF_01     5700   8989    93.0    2310    1049   16548   23935  2.99e8       0
## 4 CSF_01     5792   7602    91.8    3659    1517   22742   25079  3.04e8       0
## 5 CSF_01     5929   6962    79.4    3095    1319   16182   24309  2.62e8       0
## 6 CSF_02     7921   4813    93.8    3998    1615   30334   23910  3.16e8       0
## 7 CSF_02     7974  10196    89.6    3454    1565   12189   25758  2.31e8       0
## # ... with 16 more variables: Q30_RNA_read <dbl>, Q30_UMI <dbl>,
## #   Q30_barcodes <dbl>, Confidently_mapped_antisense <dbl>,
## #   Confidently_mapped_to_exonic_regions <dbl>,
## #   Confidently_mapped_to_genome <dbl>,
## #   Confidently_mapped_to_intergenic_regions <dbl>,
## #   Confidently_mapped_to_intronic_regions <dbl>,
## #   Confidently_mapped_to_transcriptome <dbl>, ...

## [1] "VDJ-T QC summary table"

## # A tibble: 5 x 23
##   Subpro~1 GemID Cells~2 Cells~3 Cells~4 Cells~5 Estim~6 Media~7 Media~8 Numbe~9
##   <chr>    <chr>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>
## 1 CSF_01   4608     65.3    93.4    58.6    58.6     815       2       4     478
## 2 CSF_01   4839     80.6    97.0    77.6    77.6    1912       3       7    1484
## 3 CSF_01   5700     68.9    96.0    64.8    64.8    5719       3       8    3708
## 4 CSF_01   5792     72.6    96.5    69.1    69.1    2729       3       7    1886
## 5 CSF_01   5929     83.2    97.6    80.8    80.8    4297       5      10    3471
## # ... with 13 more variables: Paired_clonotype_diversity <dbl>,
## #   Number_of_reads <dbl>, Number_of_short_reads_skipped <dbl>,
## #   Q30_RNA_read <dbl>, Q30_UMI <dbl>, Q30_barcodes <dbl>,
## #   Fraction_reads_in_cells <dbl>, Mean_reads_per_cell <dbl>,
## #   Mean_used_reads_per_cell <dbl>, Reads_mapped_to_TRA <dbl>,
## #   Reads_mapped_to_TRB <dbl>, Reads_mapped_to_any_V_D_J_gene <dbl>,
## #   Valid_barcodes <dbl>, and abbreviated variable names 1: Subproject, ...

## [1] "VDJ-B QC summary table"

## # A tibble: 2 x 27
##   Subpro~1 GemID Cells~2 Cells~3 Cells~4 Cells~5 Cells~6 Cells~7 Estim~8 Media~9
##   <chr>    <chr>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>
## 1 CSF_01   4839    100      61.5    38.5    61.5    38.5   100        13   1995 
## 2 CSF_01   5929     93.5    67.4    33.1    63.0    30.9    93.5     414   2186.
## # ... with 17 more variables: Median_IGK_UMIs_per_Cell <dbl>,
## #   Median_IGL_UMIs_per_Cell <dbl>,
## #   `Number_of_cells_with_productive_V-J_spanning_pair` <dbl>,
## #   Paired_clonotype_diversity <dbl>, Number_of_reads <dbl>,
## #   Number_of_short_reads_skipped <dbl>, Q30_RNA_read <dbl>, Q30_UMI <dbl>,
## #   Q30_barcodes <dbl>, Fraction_reads_in_cells <dbl>,
## #   Mean_reads_per_cell <dbl>, Mean_used_reads_per_cell <dbl>, ...

Back to top

## R version 4.2.2 (2022-10-31)
## Platform: x86_64-conda-linux-gnu (64-bit)
## Running under: CentOS Linux 7 (Core)
## 
## Matrix products: default
## BLAS/LAPACK: /scratch/groups/singlecell/software/anaconda3/envs/csf/lib/libopenblasp-r0.3.21.so
## 
## locale:
## [1] C
## 
## attached base packages:
## [1] stats     graphics  grDevices utils     datasets  methods   base     
## 
## other attached packages:
##  [1] RColorBrewer_1.1-3 magrittr_2.0.3     glue_1.6.2         ggrepel_0.9.3     
##  [5] ggpubr_0.6.0       gt_0.8.0           forcats_0.5.2      stringr_1.5.0     
##  [9] dplyr_1.0.10       purrr_1.0.1        readr_2.1.3        tidyr_1.3.0       
## [13] tibble_3.1.8       ggplot2_3.4.0      tidyverse_1.3.2    rmarkdown_2.20    
## 
## loaded via a namespace (and not attached):
##  [1] fs_1.6.0            lubridate_1.9.0     httr_1.4.4         
##  [4] tools_4.2.2         backports_1.4.1     bslib_0.4.2        
##  [7] utf8_1.2.2          R6_2.5.1            DBI_1.1.3          
## [10] colorspace_2.1-0    withr_2.5.0         tidyselect_1.2.0   
## [13] processx_3.8.0      compiler_4.2.2      cli_3.6.0          
## [16] rvest_1.0.3         xml2_1.3.3          labeling_0.4.2     
## [19] sass_0.4.5          scales_1.2.1        callr_3.7.3        
## [22] commonmark_1.8.1    digest_0.6.31       dichromat_2.0-0.1  
## [25] pkgconfig_2.0.3     htmltools_0.5.4     dbplyr_2.3.0       
## [28] fastmap_1.1.0       highr_0.10          maps_3.4.1         
## [31] rlang_1.0.6         readxl_1.4.1        pals_1.7           
## [34] jquerylib_0.1.4     generics_0.1.3      farver_2.1.1       
## [37] jsonlite_1.8.4      car_3.1-2           googlesheets4_1.0.1
## [40] patchwork_1.1.2     Rcpp_1.0.10         munsell_0.5.0      
## [43] fansi_1.0.4         abind_1.4-5         lifecycle_1.0.3    
## [46] stringi_1.7.12      yaml_2.3.7          carData_3.0-5      
## [49] grid_4.2.2          promises_1.2.0.1    crayon_1.5.2       
## [52] cowplot_1.1.1       haven_2.5.1         mapproj_1.2.11     
## [55] hms_1.1.2           knitr_1.41          ps_1.7.2           
## [58] pillar_1.8.1        ggsignif_0.6.4      reprex_2.0.2       
## [61] evaluate_0.20       getPass_0.2-2       modelr_0.1.10      
## [64] vctrs_0.5.2         tzdb_0.3.0          httpuv_1.6.8       
## [67] cellranger_1.1.0    gtable_0.3.1        assertthat_0.2.1   
## [70] cachem_1.0.6        xfun_0.36           broom_1.0.2        
## [73] rstatix_0.7.2       later_1.3.0         googledrive_2.0.0  
## [76] gargle_1.2.1        workflowr_1.7.0     timechange_0.2.0   
## [79] ellipsis_0.3.2